体素内不相干运动扩散加权成像定量参数评估胰腺癌的分化程度

Intra voxel Incoherent Motion Diffusion-weighted Imaging Quantitative Parameters in Evaluating Differentiated Degrees of Pancreatic Cancer

目的胰腺癌早期诊断缺乏有效的筛查工具,本研究通过对比不同病理级别胰腺癌的体素内不相干运动扩散加权成像(IVIM-DWI)表现,探讨IVIM-DWI定量参数在胰腺癌病理分级评估中的应用价值。资料与方法收集病理证实的胰腺癌患者16例(高-中分化10例,低分化6例),采用3.0T MR扫描仪行胰腺多b值DWI。应用IVIM双指数模型分析多b值DWI测量参数,测量并比较高-中分化胰腺癌和低分化胰腺癌的慢表观扩散系数(ADC_(slow))、快表观扩散系数(ADC_(fast))和灌注分数(f)。结果高-中分化胰腺癌的ADC_(slow)值低于低分化胰腺癌的ADC_(slow)值[(0.546±0.041)×10^(-3) mm^2/s比(0.677±0.120)×10^(-3) mm^2/s,P<0.05]。高-中分化胰腺癌的f值高于低分化胰腺癌的f值[(59.3±8.8)%比(41.7±22.4)%,P<0.05]。鉴别高-中分化胰腺癌和低分化胰腺癌时,ADC_(slow)的曲线下面积大于f的曲线下面积(0.850>0.750)。当ADC_(slow)值≤0.599×10^(-3) mm^2/s,鉴别高-中分化胰腺癌和低分化胰腺癌的敏感度为100.00%,特异度为83.33%。当f值>44.7%时,鉴别高-中分化胰腺癌和低分化胰腺癌的敏感度为100.00%,特异度为66.67%。结论 IVIM-DWI定量参数ADC_(slow)值和f值能够鉴别高-中分化胰腺癌和低分化胰腺癌,能够术前预测胰腺癌病理分级。鉴别高-中分化胰腺癌和低分化胰腺癌时,ADC_(slow)值和f值具有较高的诊断效能。

Purpose To explore the application of intra voxel incoherent motion diffusion-weighted imaging（IVIM-DWI） quantitative parameters in evaluating the pathological stage of pancreatic cancer by comparing the manifestations of IVIM-DWI in patients with pancreatic cancer in different differentiaed degrees as there lacked effective screening instrument for the early diagnosis of pancreatic cancer. Materials and Methods Sixteen patients with pathologically proved pancreatic cancer（10 with high-moderation differentiation while 6 with low differentiation） were enrolled, and 3.0T MRI was used to conduct pancreatic DWI with multiple b values. IVIM double-exponential model was used to analyze the measurement parameters of DWI with multiple b values, so as to measure the slow apparent diffusion coefficient（ADCslow）, fast apparent diffusion coefficient（ADCfast） and filling fraction（f）. Results The ADCslow value was evidently lower in patients with high-moderate differentiated pancreatic cancer than those with low differentiated pancreatic cancer [（0.546±0.041）×10^-3 mm^2/s vs.（0.677±0.120）×10^-3 mm^2/s, P0.05], and f value was notably higher in patients with high-moderate differentiated pancreatic cancer than those with low differentiated pancreatic cancer [（59.3±8.8）% vs.（41.7±22.4）%, P0.05]. The area under the curve of ADCslow was higher than that of f when distinguishing highmoderate differentiated and low differentiated pancreatic cancer（0.8500.750）. The sensitivity and specificity were 100.00% and 83.33% when ADCslow≤ 0.599×10^-3 mm^2/s, and were 100.00% and 66.67% when f44.7% in distinguishing high-moderate differentiated and low differentiated pancreatic cancer, respectively. Conclusion ADCslow and f, as the quantitative parameters for IVIM-DWI, can distinguish high-moderate differentiated and low differentiated pancreatic cancer, and predict the pathological stage of pancreatic cancer before operation. Moreover, they also have high diagnostic efficacy in distinguishing high-moderate differentiated and low differentiated pancreatic cancer.